ABSTRACT
The performance aspects and related technology issues of new applications are frequently not addressed in any level of detail until just about the time that they are due to go into production. The reasons that are given for failure to tackle the subject include lack of information and lack of time. It is often assumed (or should that be hoped) that only a modest amount of tuning or a judicious hardware upgrade will be necessary to cure any problems. Unfortunately, neither tuning nor hardware upgrades are necessarily a panacea, particularly on larger systems. The original decisions that were made early in the system lifecycle on the choice of software products, hardware selection, technical infrastructure, and application design can all result in problems that are costly to resolve, and that can take significant time to implement. To overcome any fundamental problems, performance needs to be treated as an integral part of the entire lifecycle, from bid or feasibility study, through the project lifecycle, and on to post-implementation. Tackling performance issues before implementation requires a specific set of skills, including a thorough understanding of hardware and software technologies, an appreciation of vendors’ financial status as well as their technical strategies and current product range, the experience to gauge the credibility of vendor claims, the confidence and ability to size the hardware when there is limited information to work with, and the capability to assess the effectiveness of proposed design techniques. In essence, the emphasis should be on pre-empting problems, rather than waiting for them to surface.
